CAREER: Processing and Optimization of Generalized Quantification

University Of Louisville Research Foundation Inc, Louisville KY

Investigators

Abstract

SQL is the standard query language for relational databases; however, it makes expression of certain queries that appear to be simple in natural language (e.g., Find the employees working in all the projects, or half the projects, or check whether all employees are working on the same number of projects) awkward or impossible. Even when such queries can be written, the optimization of their processing is very difficult. This proposal addresses this shortcoming by defining and studying a new query language, QLGQ that can express a large class of queries in a natural and easy way. Although it is more flexible and powerful than SQL, QLGQ shares important characteristics with it. Hence, it is expected that study of QLGQ and its optimization will yield insights that can be applied to SQL. The most immediate results are expected in possible extensions to SQL and optimization techniques for them. As the research extends to other data models (XML), contributions to the design of query languages for those models are also expected. The broader impacts of this proposal are significant: as more and more scientific and technological fields depend on the analysis of large amounts of accumulated data for their progress, powerful query languages that make such analysis possible become more and more important. The PI plans to explore applications of the technology by collaborating with interested potential users. It is also planned to seek commercial developments. Dissemination of all results (papers, software, presentations, experiments, etc.) is an important goal of the project and will be accomplished via a Web site http://date.spd.louisville.edu/Research/GQs.htm. The PI will also present results at workshops, talks, etc. for researchers (specialists or non-specialists), database or XML practitioners, and students. The educational plan for this project includes incorporation of the research into the PI's courses, conducting seminars with students and actively trying to recruit both graduate and undergraduate students to participate in the project.